Github user manishgupta88 commented on a diff in the pull request:
https://github.com/apache/carbondata/pull/1972#discussion_r168700461
--- Diff:
integration/spark2/src/main/scala/org/apache/spark/sql/execution/command/management/CarbonLoadDataCommand.scala
---
@@ -544,17 +547,76 @@ case class CarbonLoadDataCommand(
CarbonLoadOptionConstants.CARBON_OPTIONS_IS_EMPTY_DATA_BAD_RECORD,
isEmptyBadRecord)
CarbonSession.threadSet("partition.operationcontext", operationContext)
+ // input data from csv files. Convert to logical plan
--- End diff --
In all the above lines wherever threadset is used, everytime we are cloning
a new object for CarbonSessionInfo. As it is an internal call we can write a
overloaded method and pass boolean to avoid creating a new object everytime
---
